Results from instrumented gait analysis vary between test situations. Subject characteristics and the biomechanical model can influence the total amount of variability. The purpose of this study was to quantify reliability of gait data in general, and with respect to the applied model, and investigated population group. Reliability was compared between a functional and a predictive gait model in subjects with knee osteoarthritis and healthy controls. Day-to-day consistency for sagittal plane variables was comparable between models and population groups. Transversal plane variables relative to joint excursion showed larger inconsistency for repeated measures, even for a more sophisticated biomechanical approach. In conclusion, the presented reliability data of sagittal plane kinematics should be used for a reasonable interpretation of results derived in clinical gait analysis. Variables of the transversal plane should not be used as long as sources of error are not sufficiently minimized.
Inga Krauss, Thomas Ukelo, Christoph Ziegler, Detlef Axmann, Stefan Grau, Thomas Horstmann and Alex Stacoff
Christian Maiwald, Stefan Grau, Inga Krauss, Marlene Mauch, Detlef Axmann and Thomas Horstmann
The aim of this study was to provide detailed information on rationales, calculations, and results of common methods used to quantify reproducibility in plantar pressure variables. Recreational runners (N = 95) performed multiple barefoot running trials in a laboratory setup, and pressure variables were analyzed in nine distinct subareas of the foot. Reproducibility was assessed by calculating intraclass correlation coefficients (ICC) and the root mean square error (RMSE). Intraclass correlation coefficients ranged from 0.58 to 0.99, depending on the respective variable and type of ICC. Root mean square errors ranged between 2.3 and 3.1% for relative force–time integrals, between 0.07 and 0.23 for maximum force (Fmax), and between 107 and 278 kPa for maximum pressure (Pmax), depending on the subarea of the foot. Force–time integral variables demonstrated the best within-subject reproducibility. Rear-foot data suffered from slightly increased measurement error and reduced reproducibility compared with the forefoot.